CN112766683A - Food enterprise credit evaluation method and device and electronic equipment - Google Patents

Food enterprise credit evaluation method and device and electronic equipment Download PDF

Info

Publication number
CN112766683A
CN112766683A CN202110032792.0A CN202110032792A CN112766683A CN 112766683 A CN112766683 A CN 112766683A CN 202110032792 A CN202110032792 A CN 202110032792A CN 112766683 A CN112766683 A CN 112766683A
Authority
CN
China
Prior art keywords
training
food
enterprise
matrix
gcn
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110032792.0A
Other languages
Chinese (zh)
Inventor
张曙华
杨安荣
邬旭栋
马睿涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Xinlian Information Development Co Ltd
Original Assignee
Shanghai Xinlian Information Development Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Xinlian Information Development Co Ltd filed Critical Shanghai Xinlian Information Development Co Ltd
Priority to CN202110032792.0A priority Critical patent/CN112766683A/en
Publication of CN112766683A publication Critical patent/CN112766683A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • G06Q50/265—Personal security, identity or safety

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Marketing (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • General Business, Economics & Management (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Evolutionary Biology (AREA)
  • Computer Security & Cryptography (AREA)
  • Primary Health Care (AREA)
  • General Preparation And Processing Of Foods (AREA)

Abstract

The invention provides a food enterprise credit evaluation method, a device and electronic equipment. Wherein, the method comprises the following steps: acquiring characteristic information of a food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter; inputting the characteristic information into a Graph Convolution Network (GCN) trained in advance so that the GCN outputs a credit value of a food enterprise to be evaluated according to the characteristic information; wherein, the GCN is obtained by training based on food enterprise relations in a food supply chain; therefore, the food enterprises can be evaluated according to the credit value of the food enterprises through the GCN, the food enterprises to be judged can be comprehensively evaluated based on the relation among the food enterprises in the supply chain in the credit value evaluation process of the food enterprises, and the accuracy of the credit value of the food enterprises is improved compared with the evaluation of a single food enterprise in the existing method.

Description

Food enterprise credit evaluation method and device and electronic equipment
Technical Field
The invention relates to the technical field of enterprise credit, in particular to a food enterprise credit evaluation method, a device and electronic equipment.
Background
"the people eat as the day and eat as the first. In recent years, with the development of economy, the living standard of people is continuously improved, and the food safety problem is more and more concerned and valued by people. The current situation of food safety is still severe, and the problem of food safety is frequent, so that the establishment of a daily supervision and early warning mechanism of food enterprises has important significance.
The existing method mainly utilizes fuzzy integration and a related algorithm to evaluate the credit rating of an enterprise, specifically, firstly, characteristic attributes used for evaluating the credit rating are determined, then an attribute value-evaluation value data pair is obtained according to the existing credit rating evaluation data of the enterprise, the credit evaluation value of the enterprise is normalized to [0,1], and finally, a fuzzy measure is calculated based on a particle swarm and gradient descent mixed algorithm so as to output the corresponding credit rating of the enterprise according to the characteristic attributes of the enterprise to be evaluated. Although this method can realize the evaluation of the credit rating of the enterprise, the characteristic attributes of each enterprise are different, so the credit evaluation value of the food enterprise obtained based on the same data is not ideal, and the actual supervision requirement of the food enterprise cannot be met.
Disclosure of Invention
In view of the above, the present invention provides a food enterprise credit evaluation method, device and electronic device to alleviate the above problems and improve the credit accuracy of food enterprises.
In a first aspect, an embodiment of the present invention provides a food enterprise credit assessment method, where the method includes: acquiring characteristic information of a food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter; inputting the characteristic information into a Graph Convolution Network (GCN) trained in advance so that the GCN outputs a credit value of a food enterprise to be evaluated according to the characteristic information; wherein, the GCN is obtained by training based on food enterprise relations in a food supply chain.
With reference to the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, wherein the training process of the GCN is as follows: acquiring a graph training sample set; wherein, the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in the training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; moreover, at least part of the training characteristic information also comprises credit marking values of corresponding training food enterprises; and inputting the graph training sample set into the original GCN for training to obtain the GCN.
With reference to the first possible implementation manner of the first aspect, an embodiment of the present invention provides a second possible implementation manner of the first aspect, where the training relationship matrix includes a degree matrix and an adjacency matrix; the degree matrix is used for representing the number of the training food enterprises which have incidence relation with the current training food enterprises; the adjacency matrix is used for representing the current training food enterprise and the quantitative representation of the relationship of the training food enterprise having the incidence relation with the current training food enterprise.
With reference to the second possible implementation manner of the first aspect, an embodiment of the present invention provides a third possible implementation manner of the first aspect, where before the step of inputting the graph training sample set to the original GCN for training, the method further includes: correcting the adjacent matrix based on a preset unit matrix to obtain a corrected adjacent matrix; wherein, the order of the unit matrix is consistent with that of the adjacent matrix; and carrying out scaling processing on the degree matrix according to a preset rule to obtain the degree matrix after scaling processing.
With reference to the third possible implementation manner of the first aspect, an embodiment of the present invention provides a fourth possible implementation manner of the first aspect, where the training relationship matrix further includes a laplacian matrix; the method further comprises the following steps: and calculating to obtain a Laplace matrix according to the adjacent matrix after the correction processing and the degree matrix after the scaling processing.
With reference to the first aspect, an embodiment of the present invention provides a fifth possible implementation manner of the first aspect, where the index parameter includes at least one of: production material condition parameters, management level parameters, overall employee quality parameters, purchasing quality control parameters, production process control parameters, sales and after-sales control parameters, output product quality parameters, quality supervision and spot check performance parameters, level parameters of main technologies in China, enterprise continuous improvement mechanism parameters, energy conservation and emission reduction parameters and automation operation parameters.
In a second aspect, an embodiment of the present invention further provides a food enterprise credit assessment apparatus, where the apparatus includes: the acquisition module is used for acquiring the characteristic information of the food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter; the evaluation module is used for inputting the characteristic information into a pre-trained graph convolution network GCN so that the GCN outputs a credit value of a food enterprise to be evaluated according to the characteristic information; wherein, the GCN is obtained by training based on food enterprise relations in a food supply chain.
With reference to the second aspect, an embodiment of the present invention provides a first possible implementation manner of the second aspect, wherein the training process of the GCN is as follows: acquiring a graph training sample set; wherein, the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in the training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; moreover, at least part of the training characteristic information also comprises credit marking values of corresponding training food enterprises; and inputting the graph training sample set into the original GCN for training to obtain the GCN.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the food enterprise credit assessment method according to the first aspect when executing the computer program.
In a fourth aspect, the embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the steps of the food enterprise credit assessment method according to the first aspect.
The embodiment of the invention has the following beneficial effects:
the embodiment of the invention provides a food enterprise credit evaluation method, a device and electronic equipment, wherein the food enterprise credit evaluation is carried out through GCN, so that the food enterprise to be judged is comprehensively evaluated based on the relation among all food enterprises in a supply chain in the credit evaluation process of the food enterprise, and the credit evaluation precision of the food enterprise is improved compared with the evaluation of a single food enterprise in the existing method.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a food enterprise credit rating method according to an embodiment of the present invention;
FIG. 2 is a flowchart of a GCN training method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a food enterprise credit rating device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Aiming at the problems that the credit evaluation value of a food enterprise obtained by the existing method is not ideal and cannot meet the actual supervision requirement of the food enterprise, the embodiment of the invention provides the credit evaluation method, the device and the electronic equipment of the food enterprise.
To facilitate understanding of the embodiment, a detailed description will be given below of a food enterprise credit rating method according to an embodiment of the present invention.
The embodiment of the invention provides a food enterprise credit evaluation method, wherein an execution main body is a server stored with pre-trained GCN (Graph connected Network). The GCN is a convolutional neural network operating on a graph, and can perform deep learning on graph data, and can achieve a good effect in processing irregular data such as a graph. The GCN can extract the characteristic information of the nodes and the relation lines in the graph by effectively utilizing the topological structure of the graph, and realize information transmission between the connected nodes by carrying out weighted average calculation on the characteristics of adjacent nodes, so that the GCN can find out the inherent detailed association of data from the association of global data, thereby describing the graph data more in detail and deeply.
Based on the server, the embodiment of the invention provides a food enterprise credit assessment method, as shown in fig. 1, the method includes the following steps:
step S102, acquiring characteristic information of a food enterprise to be evaluated;
the characteristic information comprises a plurality of index parameters and enterprise data corresponding to each index parameter; specifically, the index parameter includes at least one of: production material condition parameters, management level parameters, overall employee quality parameters, purchasing quality control parameters, production process control parameters, sales and after-sales control parameters, output product quality parameters, quality supervision and spot check performance parameters, level parameters of main technologies in China, enterprise continuous improvement mechanism parameters, energy conservation and emission reduction parameters and automation operation parameters.
Specifically, from the food supply chain, the indexes influencing the food enterprise credit mainly comprise four categories of basic element score, process control score, quality performance score and development guide score; wherein each class includes the following: (1) the base element score includes: production material condition scoring, management level scoring and whole employee quality scoring; (2) the process control scoring includes: purchasing quality control scoring, production process control scoring, selling and after-sale control scoring; (3) the quality performance score includes: outputting product quality scores, quality supervision spot check performance scores and level scores of main technical indexes in China; (4) the development-oriented scores include: the enterprise continuously improves mechanism scoring, energy-saving emission-reducing scoring and automatic operation scoring.
Therefore, according to the four major indexes in the food supply chain and the twelve fine scoring items, feature information of the enterprise to be evaluated can be obtained, and a feature vector of the enterprise to be evaluated can be constructed according to the feature information, for example, the feature vector of the enterprise to be evaluated is represented by using a dimension matrix of N × C, where N represents the number of food enterprises in the food supply chain, C represents the number of index parameters in the feature information, and when the index parameters adopt the twelve fine scoring indexes, C is 12. It should be noted that, in the feature information of the enterprise to be evaluated, the number and the type of the index parameters may be set according to actual situations, which is not limited in the embodiment of the present invention.
And step S104, inputting the characteristic information into the pre-trained GCN so that the GCN outputs the credit value of the food enterprise to be evaluated according to the characteristic information.
In practical application, the food enterprises in the food chain have their sales and supply quantities affected by other food enterprises in the upstream and downstream, so that the mutual influence among the food enterprises will also affect the credit evaluation of the food enterprises. The GCN is obtained based on food enterprise relation training in a food supply chain, the credit value of the food enterprise to be evaluated is evaluated from the perspective of a global chain according to the contact and business flow between upstream and downstream food enterprises in the food chain, and accordingly the accuracy of the credit value is improved. Note that the credit value range of the enterprise to be assessed output by the GCN is [0,1 ].
According to the food enterprise credit evaluation method provided by the embodiment of the invention, the food enterprise credit evaluation is carried out through GCN, so that the food enterprise to be judged is comprehensively evaluated based on the relation among all food enterprises in a supply chain in the food enterprise credit evaluation process, and compared with the evaluation of a single food enterprise in the existing method, the food enterprise credit evaluation method improves the credit accuracy of the food enterprise.
On the basis of the above method embodiment, an embodiment of the present invention further provides a method for training a GCN, as shown in fig. 2, the method includes the following steps:
step S202, obtaining a graph training sample set;
wherein, the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in the training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; and at least part of the training characteristic information also comprises a credit marking value of the corresponding training food enterprise.
Specifically, the training characteristic information of each training food enterprise includes a plurality of training index parameters, where the training index parameters include: production material condition parameters, management level parameters, overall employee quality parameters, purchase quality control parameters, production process control parameters, sales and after-sales control parameters, output product quality parameters, quality supervision and spot check performance parameters, level parameters of main technologies in China, enterprise continuous improvement mechanism parameters, energy-saving and emission-reduction parameters and automation operation parameters, wherein each training index parameter is also provided with a score, the score is scored by a service expert according to service experience, and the value range of each score is [0,1 ]; and acquiring training enterprise data corresponding to each training index parameter, optionally, the training enterprise data needs to cover the enterprise data of all training food enterprises in a training supply chain, including the training enterprise data of each corresponding training food enterprise from the source of planting production, to the links of processing, warehousing, transportation, merchant over-sale and the like, and at least part of the training characteristic information further includes credit marking values of the corresponding training food enterprises, wherein the credit marking values can be used for manually marking credit values of part of the training food enterprises by means of expert experience so as to train the original GCN.
Since the GCN is an algorithm of a semi-supervised scene, only part of training characteristic information needs to contain credit marking values in the training process, for example, the total amount of training food enterprises with the credit marking values does not exceed 30% of all the training food enterprises in a training supply chain, so that the training difficulty caused by a large amount of training data is avoided.
In addition, the graph structure corresponding to the original graph training sample set comprises a plurality of nodes used for representing the training food enterprises, relationship lines are connected among the nodes with the association relations, and the graph can be divided into an undirected graph and a directed graph according to the fact whether the connection lines among the nodes in the graph have directions or not. In an actual food supply chain, since the dependence of the training food enterprises is unidirectional, for example, the downstream training food enterprises depend on the upstream training food enterprises, but the business flows of the upstream and downstream training food enterprises are mutual, and the equivalent exchange between the commodities (or services) and the funds is bidirectional, the graph structure constructed based on the original GCN is an undirected graph.
In the training process of the original GCN, a graph embedding technology is also adopted; in particular, graph embedding is a technology for vectorizing complex and high-dimensional graph data by using node attributes and connection relations among nodes in a graph. Graph embedding can represent nodes in a graph in a low-dimensional, real-valued, dense vector form, so that the resulting vector form can have representation and reasoning capabilities in vector space.
In practical application, when a graph is embedded, several matrixes are constructed according to topological relations in the graph, namely a degree matrix D, an adjacent matrix A and a Laplace matrix L, and the rows and the columns of the three matrixes are the number of nodes in the graph. Wherein, only the main diagonal elements in the degree matrix D have values, which represent the number of the corresponding nodes connected with other nodes, and the other elements are 0; in the adjacency matrix A, the main diagonal element is 0, if the values of the other elements are 1, the connection relationship between the corresponding row node and the column node is shown, and if the values of the other elements are 0, the connection relationship without the node is shown; the laplacian matrix L is information for characterizing the degree matrix D and the adjacency matrix a, and specifically, is obtained by subtracting the corresponding position element value in the adjacency matrix a from each element value in the degree matrix D, that is, L ═ D-a, where the degree matrix D, the adjacency matrix a, and the laplacian matrix L are different for different types of graphs.
Therefore, in the training process, the training relation matrix comprises a degree matrix and an adjacency matrix; the degree matrix is used for representing the number of the training food enterprises having incidence relation with the current training food enterprise, namely the number of other nodes connected with the node corresponding to the current training food enterprise; the adjacency matrix is used for representing the current training food enterprise and the quantitative representation of the relationship of the training food enterprise having the incidence relation with the current training food enterprise. Since the graph structure is an undirected graph, the adjacent matrix is a symmetric matrix here.
Because the neural network in the original GCN is sensitive to the numerical value representation range, the embodiment of the present invention also preprocesses the training relationship matrix before training the original GCN. Specifically, the method includes performing the correction processing on the adjacent matrix and performing the scaling processing on the contrast matrix, and it should be noted that the correction processing and the scaling processing may be performed simultaneously or in a sequence, and the specific sequence may be set according to an actual situation, which is not limited in this embodiment of the present invention.
The adjacent matrix is corrected based on a preset unit matrix to obtain a corrected adjacent matrix; wherein the order of the unit matrix is consistent with the order of the adjacent matrix. For example, for the adjacency matrix a, an identity matrix I may be added to the adjacency matrix aNThe adjacent matrix A is corrected, INIs a matrix with a main diagonal of 1, and the subscript N denotes the identity matrix INThe adjacent matrix a and the adjacent matrix a are guidelines of N rows and N columns, and the specific N value may be set according to actual situations, which is not limited to be described in the embodiment of the present invention.
In practical application, since the main diagonal elements of the adjacency matrix a are all 0, when the matrices are multiplied, the result of multiple matrix multiplication operations is more and more 0 elements, that is, the more information in the original GCN is shown to be lost, so during the process of training the original GCN, it is necessary to perform a rectification process on the adjacency matrix, where the rectification process can be understood that each node (i.e., a training food enterprise) in the original GCN has an association relationship with itself, where the rectification formula of the adjacency matrix a is as follows:
Figure BDA0002891685020000101
wherein the content of the first and second substances,
Figure BDA0002891685020000102
denotes an adjacency matrix after correction processing, A denotes an adjacency matrix, INRepresenting an identity matrix.
And carrying out scaling processing on the degree matrix according to a preset rule to obtain the degree matrix after scaling processing. Specifically, in order to eliminate the error caused by different degrees of the nodes in the original GCN during information transfer calculation, the degree matrix D needs to be scaled, for example, by inverting to perform scaling, if scaling is not performed, the node with a smaller degree is greatly affected by the characteristics of its neighbor nodes, and the node with a larger degree is affected by the result of averaging the characteristics of all neighbor nodes, so that when all nodes in the original GCN are modeled and trained uniformly, additional errors caused by different nodes due to the influence of different numbers of neighbor nodes will be caused, and the accuracy of the final GCN will be affected.
There are various methods for scaling the degree matrix D, for example, the scaling may be performed according to the following formula:
Figure BDA0002891685020000103
the scaling process is performed by dividing D-1Splitting into two D-That is, the elements on the main diagonal of the degree matrix D are inverted and then squared to form a scaled degree matrix.
In addition, the training relationship matrix further comprises a laplacian matrix; the method further comprises the following steps: and calculating to obtain a Laplace matrix according to the adjacent matrix after the correction processing and the degree matrix after the scaling processing. The specific laplace matrix can be calculated according to the following formula:
Figure BDA0002891685020000104
wherein L represents a Laplace matrix, D-The degree matrix after the scaling process is represented,
Figure BDA0002891685020000105
representing the adjacency matrix after the rectification process.
Therefore, the Laplace matrix is constructed by multiplying the left side and the right side of the adjacent matrix after the correction processing by the degree matrix after the scaling processing, so that the adjacent matrix after the correction processing is subjected to row normalization scaling processing and column normalization scaling processing at the same time, and the Laplace matrix obtained through calculation not only can accurately represent the graph relation in the original GCN, but also can be converged more easily during model training.
In addition, the original GCN is different from other machine learning algorithms, and in addition to finding the mapping relationship from the feature vector to the target value to be predicted, the dependency relationship between data in the non-euclidean space is brought into modeling, that is, the number of other training food enterprises in the training supply chain having an association relationship with the current training food enterprise.
Step S204, inputting the graph training sample set into the original GCN for training to obtain the GCN.
In view of the advantages of GCN modeling and information transfer from a global perspective, GCN stacked networks do not need to be too deep to achieve fast convergence, for example, a two-layer GCN stacked network is used; meanwhile, the credit marking value does not need to cover all training food enterprises, under a semi-supervised scene, model training only utilizes forward propagation to calculate loss and backward propagation to calculate gradient for the training food enterprises marked with the enterprise credit, and the original GCN is iterated for multiple times until the loss function of the network is converged or the parameter value is not changed any more, so that the network training is finished.
Since the prediction of credit values for food enterprises is a regression problem, the two-layer original GCN is taken as an example here, and specifically, the loss function of the original GCN can be calculated according to the following formula:
Z＝f(X,L)＝MSE(L·ReLU(L·X·W0)·W1,y) (4)
wherein, L represents a Laplace matrix, MSE represents a loss function, X represents a feature vector corresponding to feature information, which can be understood as a feature vector corresponding to enterprise data of all training food enterprises in a training supply chain, and W represents a feature vector corresponding to enterprise data of all training food enterprises in a training supply chain0And W1Representing the network parameters of the first layer and the second layer in the original GCN, y represents the credit mark value, therefore, the calculation of the loss function can be applied to all the training food enterprises with the credit mark value in the training supply chainAnd solving the minimum mean square error of the credit label value y and the predicted value y ', namely calculating the MSE (y', y). When the loss function converges, a trained GCN is obtained, and the credit value of the food enterprise to be judged can be obtained through the GCN. It should be noted that, in the training process, the score range of each training index parameter is [0,1]]Therefore, the credit value range of the food enterprise to be judged, which is output by the GCN in practical application, meets the requirement of [0, 1%]。
Therefore, the food enterprise credit assessment method provided by the embodiment of the invention has the following advantages: (1) in the process of evaluating enterprise credit, the original GCN is modeled through the topological relation of food enterprises among supply chains from the perspective of a full supply chain, and the original GCN is not predicted one by one, so that the enterprise credit score is predicted more accurately, and overfitting is not easy to occur; (2) in the training process, a small amount of marking data is used for training an original GCN model, so that the manual marking cost is reduced, and the precision is improved on the premise of not using a pre-training model; (3) the model training is easier to be converged, thereby being beneficial to accelerating the model iteration and updating the enterprise credit in time; (4) in the prediction process, the food enterprise relation does not need to be reconstructed, and only the new food enterprise relation and other food enterprise relations need to be added in the existing graph structure relation, so that the characteristic information of the food enterprise to be evaluated is input, the credit value of the food enterprise to be evaluated is output, the credit iteration of the supply chain enterprise is facilitated, and the method has a good practical value.
On the basis of the above method embodiment, the embodiment of the present invention further provides a food enterprise credit assessment apparatus, as shown in fig. 3, the apparatus includes an obtaining module 31 and an assessment module 32; the functions of each module are as follows:
the obtaining module 31 is used for obtaining the characteristic information of the food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter;
the evaluation module 32 is used for inputting the characteristic information into a graph convolution network GCN trained in advance so that the GCN outputs a credit value of a food enterprise to be evaluated according to the characteristic information; wherein, the GCN is obtained by training based on food enterprise relations in a food supply chain.
The food enterprise credit evaluation device provided by the embodiment of the invention evaluates the credit value of the food enterprise through the GCN, realizes the comprehensive evaluation of the food enterprise to be judged based on the relation among all the food enterprises in the supply chain in the credit value evaluation process of the food enterprise, and improves the credit value precision of the food enterprise compared with the evaluation of a single food enterprise in the existing method.
In one possible embodiment, the training process of the GCN is as follows: acquiring a graph training sample set; wherein, the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in the training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; moreover, at least part of the training characteristic information also comprises credit marking values of corresponding training food enterprises; and inputting the graph training sample set into the original GCN for training to obtain the GCN.
In another possible embodiment, the training relationship matrix includes a degree matrix and an adjacency matrix; the degree matrix is used for representing the number of the training food enterprises which have incidence relation with the current training food enterprises; the adjacency matrix is used for representing the current training food enterprise and the quantitative representation of the relationship of the training food enterprise having the incidence relation with the current training food enterprise.
In another possible embodiment, before inputting the graph training sample set into the original GCN for training, the apparatus further includes: correcting the adjacent matrix based on a preset unit matrix to obtain a corrected adjacent matrix; wherein, the order of the unit matrix is consistent with that of the adjacent matrix; and carrying out scaling processing on the degree matrix according to a preset rule to obtain the degree matrix after scaling processing.
In another possible embodiment, the training relationship matrix further includes a laplacian matrix; the device also includes: and calculating to obtain a Laplace matrix according to the adjacent matrix after the correction processing and the degree matrix after the scaling processing.
In another possible embodiment, the index parameter includes at least one of: production material condition parameters, management level parameters, overall employee quality parameters, purchasing quality control parameters, production process control parameters, sales and after-sales control parameters, output product quality parameters, quality supervision and spot check performance parameters, level parameters of main technologies in China, enterprise continuous improvement mechanism parameters, energy conservation and emission reduction parameters and automation operation parameters.
The food enterprise credit evaluation device provided by the embodiment of the invention has the same technical characteristics as the food enterprise credit evaluation method provided by the embodiment, so that the same technical problems can be solved, and the same technical effects can be achieved.
The embodiment of the invention also provides electronic equipment which comprises a processor and a memory, wherein the memory stores machine executable instructions capable of being executed by the processor, and the processor executes the machine executable instructions to realize the food enterprise credit assessment method.
Referring to fig. 4, the electronic device includes a processor 40 and a memory 41, the memory 41 stores machine executable instructions capable of being executed by the processor 40, and the processor 40 executes the machine executable instructions to implement the food enterprise credit assessment method.
Further, the electronic device shown in fig. 4 further includes a bus 42 and a communication interface 43, and the processor 40, the communication interface 43 and the memory 41 are connected through the bus 42.
The Memory 41 may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 43 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, etc. may be used. The bus 42 may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, an EISA (Enhanced Industry Standard Architecture) bus, or the like. The above-mentioned bus may be classified into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one double-headed arrow is shown in FIG. 4, but that does not indicate only one bus or one type of bus.
The processor 40 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 40. The Processor 40 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory 41, and the processor 40 reads the information in the memory 41 and completes the steps of the method of the foregoing embodiment in combination with the hardware thereof.
The present embodiments also provide a machine-readable storage medium having stored thereon machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the food enterprise credit assessment method described above.
The food enterprise credit assessment method, the food enterprise credit assessment device and the computer program product of the electronic device provided by the embodiment of the invention comprise a computer readable storage medium storing program codes, wherein instructions included in the program codes can be used for executing the method described in the previous method embodiment, and specific implementation can refer to the method embodiment, and is not described herein again.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In addition, in the description of the embodiments of the present invention, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (10)

1. A food enterprise credit rating method, comprising:
acquiring characteristic information of a food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter;
inputting the characteristic information into a Graph Convolution Network (GCN) trained in advance so that the GCN outputs a credit value of the food enterprise to be assessed according to the characteristic information; wherein the GCN is trained based on food business relationships in a food supply chain.
2. The food service enterprise credit rating method of claim 1, wherein the training process of the GCN is as follows:
acquiring a graph training sample set; wherein the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in a training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; moreover, at least part of the training characteristic information also comprises a credit marking value of the corresponding training food enterprise;
and inputting the graph training sample set into an original GCN for training to obtain the GCN.
3. The food enterprise credit rating method of claim 2, wherein the training relationship matrix comprises a degree matrix and an adjacency matrix; wherein the degree matrix is used for representing the number of the training food enterprises which have incidence relation with the current training food enterprises; the adjacency matrix is used for representing the current training food enterprise and the quantitative representation of the training food enterprise relationship having the incidence relation with the current training food enterprise.
4. The food service establishment credit rating method of claim 3, wherein prior to the step of inputting the set of graph training samples into an original GCN for training, the method further comprises:
correcting the adjacent matrix based on a preset unit matrix to obtain the corrected adjacent matrix; wherein the order of the unit matrix is consistent with the order of the adjacent matrix;
and carrying out scaling processing on the degree matrix according to a preset rule to obtain the degree matrix after scaling processing.
5. The food service establishment credit assessment method according to claim 4, wherein said training relationship matrix further comprises a Laplace matrix; the method further comprises the following steps:
and calculating to obtain the Laplace matrix according to the adjacent matrix after the correction processing and the degree matrix after the scaling processing.
6. The food enterprise credit rating method of claim 1, wherein the indicator parameter comprises at least one of: production material condition parameters, management level parameters, overall employee quality parameters, purchasing quality control parameters, production process control parameters, sales and after-sales control parameters, output product quality parameters, quality supervision and spot check performance parameters, level parameters of main technologies in China, enterprise continuous improvement mechanism parameters, energy conservation and emission reduction parameters and automation operation parameters.
7. A food enterprise credit rating device, said device comprising:
the acquisition module is used for acquiring the characteristic information of the food enterprise to be evaluated; the characteristic information comprises index parameters and enterprise data corresponding to each index parameter;
the evaluation module is used for inputting the characteristic information into a pre-trained Graph Convolution Network (GCN) so that the GCN outputs a credit value of the food enterprise to be evaluated according to the characteristic information; wherein the GCN is trained based on food business relationships in a food supply chain.
8. The food service establishment credit rating device of claim 7, wherein the GCN training process is as follows:
acquiring a graph training sample set; wherein the graph training sample set comprises: training a relation matrix and a training feature set; the training relation matrix is used for representing the incidence relation of each training food enterprise in a training supply chain; the training feature set comprises training feature information of each training food enterprise; the training characteristic information comprises training index parameters and training enterprise data corresponding to each training index parameter; moreover, at least part of the training characteristic information also comprises a credit marking value of the corresponding training food enterprise;
and inputting the graph training sample set into an original GCN for training to obtain the GCN.
9. An electronic device comprising a memory, a processor and a computer program stored on said memory and executable on said processor, wherein said processor when executing said computer program performs the steps of the food industry credit assessment method according to any of the preceding claims 1-6.
10. A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the food service establishment credit rating method according to any one of claims 1 to 6 are performed.
CN202110032792.0A 2021-01-11 2021-01-11 Food enterprise credit evaluation method and device and electronic equipment Pending CN112766683A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110032792.0A CN112766683A (en) 2021-01-11 2021-01-11 Food enterprise credit evaluation method and device and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110032792.0A CN112766683A (en) 2021-01-11 2021-01-11 Food enterprise credit evaluation method and device and electronic equipment

Publications (1)

Publication Number Publication Date
CN112766683A true CN112766683A (en) 2021-05-07

Family

ID=75701374

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110032792.0A Pending CN112766683A (en) 2021-01-11 2021-01-11 Food enterprise credit evaluation method and device and electronic equipment

Country Status (1)

Country Link
CN (1) CN112766683A (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107392456A (en) * 2017-07-14 2017-11-24 武汉理工大学 A kind of multi-angle rating business credit modeling method for merging internet information
CN109670944A (en) * 2018-12-19 2019-04-23 信雅达系统工程股份有限公司 A kind of rating business credit method and system based on map relational network
CN111222681A (en) * 2019-11-05 2020-06-02 量子数聚（北京）科技有限公司 Data processing method, device, equipment and storage medium for enterprise bankruptcy risk prediction
US20200219034A1 (en) * 2019-01-03 2020-07-09 International Business Machines Corporation Quality score for a food supply chain
US20200285944A1 (en) * 2019-03-08 2020-09-10 Adobe Inc. Graph convolutional networks with motif-based attention
CN111859166A (en) * 2020-07-28 2020-10-30 重庆邮电大学 Article scoring prediction method based on improved graph convolution neural network

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107392456A (en) * 2017-07-14 2017-11-24 武汉理工大学 A kind of multi-angle rating business credit modeling method for merging internet information
CN109670944A (en) * 2018-12-19 2019-04-23 信雅达系统工程股份有限公司 A kind of rating business credit method and system based on map relational network
US20200219034A1 (en) * 2019-01-03 2020-07-09 International Business Machines Corporation Quality score for a food supply chain
US20200285944A1 (en) * 2019-03-08 2020-09-10 Adobe Inc. Graph convolutional networks with motif-based attention
CN111222681A (en) * 2019-11-05 2020-06-02 量子数聚（北京）科技有限公司 Data processing method, device, equipment and storage medium for enterprise bankruptcy risk prediction
CN111859166A (en) * 2020-07-28 2020-10-30 重庆邮电大学 Article scoring prediction method based on improved graph convolution neural network

Similar Documents

Publication Publication Date Title
CN110060144B (en) Method for training credit model, method, device, equipment and medium for evaluating credit
CN109389151B (en) Knowledge graph processing method and device based on semi-supervised embedded representation model
CN107169628A (en) A kind of distribution network reliability evaluation method based on big data mutual information attribute reduction
CN108898476A (en) A kind of loan customer credit-graded approach and device
CN106296669A (en) A kind of image quality evaluating method and device
CN113190688B (en) Complex network link prediction method and system based on logical reasoning and graph convolution
CN111582538B (en) Community value prediction method and system based on graph neural network
CN106897703A (en) Remote Image Classification based on AGA PKF SVM
CN104462459A (en) Neural network based big data analysis and processing system and method
CN111967796A (en) Food enterprise credit scoring method and device and electronic equipment
CN110175689A (en) A kind of method of probabilistic forecasting, the method and device of model training
CN112613658A (en) Method and device for predicting rainfall day by day, electronic equipment and storage medium
CN114491263A (en) Recommendation model training method and device, and recommendation method and device
CN112215398A (en) Power consumer load prediction model establishing method, device, equipment and storage medium
CN114817571A (en) Method, medium, and apparatus for predicting achievement quoted amount based on dynamic knowledge graph
CN113723541B (en) Slope displacement prediction method based on hybrid intelligent algorithm
CN113610220A (en) Training method, application method and device of neural network model
Wang et al. An iterative updating heuristic search strategy for groundwater contamination source identification based on an ACPSO–ELM surrogate system
CN111814804B (en) Human body three-dimensional size information prediction method and device based on GA-BP-MC neural network
CN109697531A (en) A kind of logistics park-hinterland Forecast of Logistics Demand method
CN112766683A (en) Food enterprise credit evaluation method and device and electronic equipment
CN108009859A (en) Agricultural product price fluctuates method for early warning and equipment
CN113393023B (en) Mold quality evaluation method, apparatus, device and storage medium
CN114066081B (en) Enterprise risk prediction method and device based on graph attention network and electronic equipment
CN112766684A (en) Enterprise credit evaluation method and device and electronic equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination